Current Research

Since the industrial revolution, the ocean has taken up ~28% of all carbon dioxide emitted by human activities (IPCC AR5). The vast majority of this uptake has occured through the "solubility pump" whereby cold high latitude waters (which can hold more gas) densify and sink into the ocean interior, bringing anthropogenic carbon with them. The solubility pump is not completely efficient, however, and its unclear how exactly its efficiency will change in the future. Because noble gases are also taken up by cooling, densifying high latitude waters but are only sensitive to physics, their concentrations throughout the ocean interior provide important clues about the physical mechanisms that regulate gas uptake efficiency during deep-water formation. Our group is working with collaborators to further develop noble gases and their isotopes as a suite of complementary tracers of the ventilation efficiency of the deep ocean. We are developing new analytical techniques, making new measurements in the deep ocean, incorporating these tracers into models, and carrying out lab experiments to better constrain their sensitivities to different processes.

Check out our group's recent study applying a new technique at WHOI to measure samples from the deep North Atlantic and constrain ventilation processes, published March 2023 in Proceedings of the National Academy of Sciences (PDF).

We are grateful to the NSF Chemical Oceanography and Antarctic Science programs for supporting this work.

Below: The annual cycle of argon isotopes in the Irminger Sea, from model simulations in our 2023 PNAS paper. The signal of isotopic disequilibrium, which is set by diffusion across bubbles and the air-sea interface in high-latitude deep water formation events is preserved downstream in the deep North Atlantic, e.g., off the coast of Bermuda where we collected and analyzed samples. 

The last glacial maximum (LGM; ~20,000 years ago) is the most recent stable, long-term climate period that was substantially colder than the present. That makes it an excellent target for evaluating climate models and estimating Earth's climate sensitivity. However, accurate and quantitative constraints on LGM surface temperature are needed, especially on land at low elevation. The "noble gas paleothermometer" enables reconstruction of mean annual surface temperature based on measurements of neon, argon, krypton, and xenon dissolved in ancient groundwater. There are two key properties of these gases that make them excellent records of past temperature: (1) each gas has a well defined temperature-dependent solubility in water, and (2) these inert gases lack any sinks or sources within an aquifer, so their composition is preserved over time.

In our 2021 paper in Nature, we compiled dozens of noble gas records in LGM groundwater to find that the low-to-mid latitudes cooled by roughly 6 degrees C.

In our 2023 paper in Science Advances, we explored the thermodynamic implications of cooling on land during the LGM within the context of "terrestrial amplification" - that is, the enhancement of warming/cooling over land, relative to oceans - using the past (i.e., the LGM) to constrain a theory to predict future warming on land.

In our 2024 paper in Geophysical Research Letters, we built upon our recent work using noble gas tracers of air-sea disequilibrium in the modern ocean to run a modern data-validated model under LGM boundary conditions to quantify and correct for the role of air-sea disequilibrium in past ice-core noble gas-based records of LGM mean ocean temperature.

We are grateful to the NSF P2C2 and Antarctic Research progams for supporting this work.

Below: An illustration of how a moist static energy framework for terrestrial amplification can be adapted to the LGM and informed by constraints on past cooling that come from noble gases in groundwater (reproduced from Seltzer et al., 2023, Sci. Adv.)

Our group has developed a new technique to measure the dissolved Kr and Xe composition of groundwater at the order 0.001 per mil amu-1 scale. At this level, we are able to reconstruct the depth to water (i.e., water table depth) at the time and place a parcel of groundwater was "recharged" (i.e., was last in contact with the atmosphere). The basis for this new paleohydrological tool is rooted in gravitational settling in soil air, a well-understood physical process that results in enrichment of heavy-to-light gas ratios with depth. As Kr and Xe in soil air dissolve into groundwater at the water table, the gravitational signal is ‘locked in’ and is preserved over time in the aquifer (Seltzer et al., 2019). 

In addition to water table depth reconstruction, we are also working to develop high-precision measurements of "triple argon isotopes" (argon-40, argon-38, and argon-36) to deconvolve physical fractionation and radiogenic accumulation of argon-40, which is produced from potassium-40 decay in aquifer minerals. We are pairing new argon isotope measurements with krypton-81 (half life = 229,000 years) and radiocarbon (half life = 5,730 years) to gain insights into groundwater residence time distributions and flow pathways, providing useful tools to inform groundwater sustainability. MIT-WHOI PhD Student Grace Brown is leading a new study to explore all of these tracers (Ar + Kr + Xe stable isotopes & krypton-81) in fossil groundwater systems from Australia, to gain insight into past climate and hydrogeology. 

We are grateful to the NSF P2C2 (now P4CLIMATE) and Hydrological Sciences programs for supporting this work.

Below: A demonstration of gravitational signals in dozens of groundwater wells in California at a scale previously well below the detection limit of conventional analytical techniques. For a full description of this figure, please check out the original publication here.

In collaboration with colleagues at WHOI and CRPG (France), our lab has developed a new technique to measure the isotopes of heavy noble gases in volcanic gas samples at several orders-of-magnitude higher precision that previously possible. This new approach holds promise for robustly disentangling signals of primary volatiles (those acquired when Earth formed), secondary volatiles (those produced within the Earth since formation), and physical processes (e.g., diffusive fractionation) that modify the observed composition.

Since our lab started in 2021, we have developed and tested our new technique (2022 IJMS method paper) and applied it to analyze new samples collected from Iceland, Yellowstone, Costa Rica, Chile, Argentina, Bolivia, Germany, Salton Sea, Italy, Djibouti, Kenya, France, New Zealand. Former postdoc David Bekaert recently led the first study applying our new technique and introducing a new conceptual framework for subsurface fractionation of volcanic gases (Bekaert et al., 2023, Sci. Adv.) - stay tuned with several more exciting new results from our lab to be released in the near future!

We are grateful to the NSF Geochemistry and Petrology program for supporting this work.

Below: An overview of the setting, sampling technique, and subsurface fractionation concept suggested by Bekaert et al. (2023), Sci. Adv.

In collaboration with several colleagues, we are working to better understand what processes give rise to new experimentally observed differences in solubility and diffusivity among isotopes of noble gases. For example in our 2019 paper, we found that the heavy isotopes of Ar, Kr, and Xe are slightly more soluble than the light isotopes, with curious patterns of isotopic mass dependence and temperature dependence. We have been running molecular dynamics (MD) simulations using LAMMPS to look at how individual noble gas atoms collide with and travel around water molecules under various conditions. By examining the trajectories (and instantaneous velocities) of noble gas isotopes of varying mass at very small scales in space and time, and using a framework originally developed for vapor pressure isotope effects to understand these solubility effects, we can compare simulated isotopic solubility differences to measurements. The results have been encouraging, suggesting that classical MD simulations can resolve equilibrium fractionation factors for monatomic gases in water, and we recently published a Journal of Physical Chemistry paper on our findings.

Below: Simulation of an Ar-40 atom (central sphere; enlarged for clarity) interacting with water molecules (red/white objects: red = oxygen, white = hydrogen) in a 30x30x30 angstrom box. For a sense of scale, note that the Ar-40 atom tends to collide with a water molecule about twice per picosecond. By comparing the rattling frequencies of different isotopes, we can estimate solubility isotope effects and compare with observations.

In addition to telling us about air-sea gas exchange, noble gas measurements and model simulations allow us to precisely predict the abundance and composition of other gases in the ocean interior expected due to physics alone. This means that by measuring nitrogen (N2) and its isotopic composition, and comparing to noble gas-based predictions, we can identify "excess" N2 produced by biologic and use isotopic constraints to pinpoint the source of this N2. In collaboration with nitrogen cycle experts at WHOI and beyond, our group has recently started to develop a new technique in the lab to precisely measure the N2/Ar and N2 isotope composition of seawater with the goal of quantifying excess N2  produced by denitrification in the water column and sediments. New constraints on the rate of benthic denitrification and its spatial distribution can inform biogeochemical models and provide insight into how the availability of fixed nitrogen in the deep ocean may change in our warming world. MIT-WHOI Ph.D. student Katelyn McPaul is developing and applying new techniques to measure N2/Ar, N2 isotopes, and N2O isotopes throughout the interior of the Atlantic Ocean using a remarkable set of archived gas samples collected and preserved by our colleague Bill Smethie (LDEO). Stay tuned for updates as this work progresses!

Below: A pilot study combining noble gas measurements, modeling, and N2/Ar measurements in archived samples throughout the deep North Atlantic reveals excess N2 in the oldest, easternmost waters (Seltzer et al., 2023, PNAS)